(1) Field of the Invention
The method disclosed herein allows for the melting of cesium germanium halides without decomposition and a direct synthesis without the introduction of possible contaminants.
(2) Description of Related Art Cesium germanium halides are a highly promising class of material for the nonlinear optical frequency conversion of laser sources. These materials promise unique capabilities for conversion from the visible to the far-infrared, with the possibility of periodic poling.
Nonlinear Optical Frequency Conversion
Nonlinear optical (NLO) frequency conversion techniques allow the wavelengths of laser sources to be changed to fit a given application. In all such techniques, the light from a pump laser is passed through an NLO material and nonlinearities in the interaction of the laser with the material give rise to new optical frequencies.
Common examples of NLO frequency converters include second-harmonic generation (SHG) and optical parametric oscillation (OPO), the latter of which is illustrated in FIG. 1.
Pump light at a frequency, ωP(ω=2π/λ), is brought into an optical cavity enclosing an NLO crystal. Light is generated at two new frequencies, the signal, ωS and the idler, ωI, where, by convention, ωS>wI. The cavity is generally designed to confine and oscillate one or more of the three frequencies involved. The NLO conversion process must satisfy conservation of energy:ωP=ωS+ωI   (1)
The process must also satisfy conservation of momentum, or phasematching. Traditional phasematching methods require satisfying Eq. 2:kP=kS−kI   (2)                where kP, kS, and kI are the wavevectors of the pump, signal, and idler respectively:k=nω  (3)        
As the indices of refraction of a material generally vary with frequency, the condition in Equation 2 is generally met by adjusting the polarizations of the three fields with respect to the optical axes of a birefringent crystal.
The efficiency of this conversion process (η) depends on the length of interaction (L), the input power (P), and the effective nonlinear coefficient of the crystal (deff):
                    η        ∝                              d            eff            2                    ⁢          LP          ⁢                                          ⁢          sin          ⁢                                          ⁢                      c            ⁡                          (                                                Δ                  ⁢                                                                          ⁢                  kL                                2                            )                                                          (        4        )                            where Δk is the phase mismatch:        
                              Δ          ⁢                                          ⁢          k                =                              k            P                    -                      k            S                    -                      k            I                                              (        5        )                                          sin          ⁢                                          ⁢                      c            ⁡                          (              x              )                                      =                              sin            ⁡                          (              x              )                                x                                    (        6        )            
So, increasing the effective nonlinear coefficient greatly increases the conversion efficiency at a given power, while increasing the pathlength increases the power at the expense of increasing the sensitivity to deviations from perfect phasematching. Birefringent walkoff, where the extraordinary ray propagates in a slightly different direction than the ordinary ray, often limits achievable interaction lengths. The dependence on pump power means that high peak power pump sources, such as Q-switched or mode-locked lasers, are often preferred over CW pumps.
Nonlinear Optical Materials
The key component of a nonlinear optical frequency converter is the NLO material. As most technologically significant NLO frequency conversion methods are based on three wave mixing processes, good NLO materials must possess a large second-order optical nonlinear susceptibility, or χ(2), where deff=χ(2)/2. For χ(2) to be greater than zero, the material must not possess inversion symmetry; this limitation severely limits the number of available materials. A good NLO material must also be transparent at all of the wavelengths involved in the conversion process and have a high damage threshold. It must be possible to phase match the desired process, so the material must either be sufficiently birefringent or made into a periodically-poled structure. Finally, the material should be as environmentally stable and robust as possible.
There are two broad classes of nonlinear optical material currently used in frequency conversion applications: oxides and semiconductors. The oxides include such common materials as potassium titanyl phosphate (KTiOPO4, or KTP), β-barium borate (β-BaB2O4, or BBO), and lithium niobate (LiNbO3). These materials are generally transparent in the visible and the near infrared (to ˜3 μm), and can be pumped with readily-available laser sources, such as Nd:YAG, near 1 μm. They can be birefringently phase matched over a large range, and periodic poling has been demonstrated in KTiOPO4 (PPKTP) and LiNbO3 (PPLN). PPLN has been a particularly successful material, finding use in many applications and enabling high-efficiency CW OPO designs.
The latter class, semiconductors, includes primarily materials developed for use in the mid-infrared at wavelengths longer than 1.5 μm, such as zinc germanium phosphide (ZnGeP2, or ZGP) and silver gallium selenide (AgGaSe2). Due to significant absorption in the visible and near-infrared, none of these materials are compatible in practice with 1 μm pump sources.
The cesium germanium halides (CsGeX3, for X=Cl, Br, I) are an emerging class of NLO materials that hold the promise of replacing both oxide and semiconductor materials in many applications. Though they were first synthesized many years ago, it is primarily since the publication of an improved synthesis and a patent on their use as nonlinear optical materials in 1998 that they have come to the attention of the nonlinear optics community.
Reports on powders have shown that they are transparent from the visible through the far-infrared, with no absorbing regions in between, and that they can be phase matched for frequency doubling of 1 μm lasers. They have also demonstrated high nonlinear coefficients (deffof ˜10 pm/V for CsGeBr3, compared with ˜4 pm/V for KTP), with evidence for the ability to birefringently phasematch the frequency doubling of lasers with wavelengths around 1 μm.
Solution Synthesis of CsGeCl3 and CsGeBr3 
All published methods for the synthesis of CsGeCl3 and CsGeBr3 are carried out in aqueous solutions. The typical reaction used involves the reduction of Ge(IV) to Ge(II) by an excess of hypophosphorous acid:H3PO2+2GeO2+4HX+2CsX→2CsGeX3+2H2O+H3PO4   (7)                where X=Cl or Br.        
This reaction is effective in generating the desired CGX, but it can leave behind residual reactants or products, and drying the CGX by conventional methods may not remove the residual water. An infrared spectrum of CsGeCl3 produced by this method is shown in FIG. 2. There are multiple spectral features present, including the phosphorous-oxygen feature at ˜1100 cm−1 and hydrogen-oxygen feature at ˜3500 cm−1. In pure CsGeCl3, there would be no absorption features at all in this region. These contaminants pose a real problem for the use of CGX compounds in practical nonlinear conversion systems, such as OPOs, as their absorption features will preclude proper production of many useful infrared wavelengths.
Thus, it is critical that uncontaminated material be produced. Purification by recrystallization (dissolving the compound in a minimal amount of hot solvent, then letting the solvent cool) has the potential to remove many of the contaminants, but residual water will continue to be a problem.
Melt Stability of CsGeCl3 and CsGeBr3 
As mentioned above, the solution synthesis route to CsGeX3 generally produces material that is contaminated with water. To remove this water, it would be desirable to purify the material via directional solidification (whereby a charge is melted and solidified slowly from one end to the other and the contaminated material that solidifies last is discarded) or zone refining. A stable melt is also required for growing single crystals from the melt.
While both cesium germanium chloride and cesium germanium bromide can be melted (with both melting around 355° C.) , they have issues with decomposition due to the relative instability of Ge(II) oxidation state compared to Ge(0) and Ge(IV). If they are melted under vacuum, under an inert atmosphere, or under GeX4, they can decompose with reduction of the Ge(II) and release of halogen gas:CsGeX3→CsX+Ge(0)+X2   (8)
They can also disproportionate:2CsGeX3→2CsX+Ge(0)+GeX4   (9)
If, instead of vacuum or an inert atmosphere, the halogen is used as a cover gas, then the Ge(II) can oxidize to Ge(IV):CsGeX3+X2→CsX+GeX4   (10)
Evidence for all of these processes have been observed for X=Cl. Under high-pressure (above 100 torr) amostpheres of Cl2, condensation of GeCl4 (indicating the action of the latter reaction, above) was evident. Under vacuum, lower-pressure Cl2, argon, or GeCl4 atmospheres, the formation of black particles was observed in the melt, indicating the presence of Ge(0) and the action of one or both of the former reactions.